Training Object Detectors from Few Weakly-Labeled and Many Unlabeled Images - Inria - Institut national de recherche en sciences et technologies du numérique
Journal Articles Pattern Recognition Year : 2021

Training Object Detectors from Few Weakly-Labeled and Many Unlabeled Images

Zhaohui Yang
  • Function : Author
  • PersonId : 1123405
Miaojing Shi
  • Function : Author
  • PersonId : 1123406
Chao Xu
Vittorio Ferrari
  • Function : Author
  • PersonId : 1123408

Abstract

Weakly-supervised object detection attempts to limit the amount of supervision by dispensing the need for bounding boxes, but still assumes image-level labels on the entire training set. In this work, we study the problem of training an object detector from one or few images with image-level labels and a larger set of completely unlabeled images. This is an extreme case of semi-supervised learning where the labeled data are not enough to bootstrap the learning of a detector. Our solution is to train a weakly-supervised student detector model from image-level pseudo-labels generated on the unlabeled set by a teacher classifier model, bootstrapped by region-level similarities to labeled images. Building upon the recent representative weakly-supervised pipeline PCL [1], our method can use more unlabeled images to achieve performance competitive or superior to many recent weakly-supervised detection solutions. Code will be made available at https://github.com/zhaohui-yang/NSOD.
Fichier principal
Vignette du fichier
J31.pr21.nsod.pdf (3.03 Mo) Télécharger le fichier
Origin Files produced by the author(s)

Dates and versions

hal-03530130 , version 1 (17-01-2022)

Identifiers

Cite

Zhaohui Yang, Miaojing Shi, Chao Xu, Vittorio Ferrari, Yannis Avrithis. Training Object Detectors from Few Weakly-Labeled and Many Unlabeled Images. Pattern Recognition, 2021, 120, pp.1-10. ⟨10.1016/j.patcog.2021.108164⟩. ⟨hal-03530130⟩
30 View
89 Download

Altmetric

Share

More